AGA Abstracts higher risk. NSAIDs with long plasma half-life and with a slow release formulation were associated with a greater risk than NSAIDs with a short half-life. Conclusions: The risk of UGIB varied between individual NSAIDs at the doses commonly used in the general popula- tion. Drugs with long half-life or slow-release formulation and/or associated with profound and coincident inhibition of both COX-isozymes were associated with a greater risk of UGIB. W1004 Time Trends in the Prescription of Low-Dose Acetylsalicylic Acid, Clopidogrel and Proton Pump Inhibitors in UK Primary Care Luis A. Garcia Rodriguez, Lucía Cea Soriano, Saga Johansson Introduction: American College of Cardiology/American Heart Association and European Society of Cardiology guidelines recommend treatment with low-dose acetylsalicylic acid (ASA) for the secondary prevention of cardiovascular events, and dual antiplatelet therapy with ASA and clopidogrel in patients with acute coronary syndromes. In addition, prophy- lactic prescription of a proton pump inhibitor (PPI) is recommended for patients who have a history of, or who are at risk of, upper gastrointestinal bleeding. The aim of this study was to assess time trends in the prescription of these drugs in UK primary care. Methods: The Health Improvement Network (THIN) UK primary care database was used to identify patients aged 50-84 years who received a first prescription of low-dose ASA (75-300 mg/ day) for secondary prevention of cardiovascular events either in 2000-2001 (N=10 330) or in 2006-2007 (N=8154). Patient records were examined to assess how many patients were prescribed clopidogrel and/or a PPI in addition to low-dose ASA. Each cohort was divided into four mutually exclusive categories, based on additional prescriptions that were received either on the day of the first low-dose ASA prescription or in the 15 days following the ASA prescription: 1) patients who received an additional prescription for clopidogrel; 2) patients who received an additional prescription for a PPI; 3) patients who received additional prescriptions for both clopidogrel and a PPI; and 4) patients who did not receive an additional prescription for clopidogrel or a PPI. Results: In 2000-2001, 87.2% (9007/10 330) of individuals were prescribed only low-dose ASA therapy; however, by 2006-2007 this propor- tion had decreased to 56.4% (4596/8154). During the study, the proportion of patients prescribed low-dose ASA and concomitant clopidogrel increased from 1.2% (126/10 330) to 15.3% (1249/8154) and the proportion of patients prescribed low-dose ASA and a PPI increased from 11.2% (1152/10 330) to 18.4% (1502/8154). The greatest relative increase over time was in the proportion of patients prescribed low-dose ASA, clopidogrel and a PPI, which increased from 0.4% in 2000-2001 (45/10 330) to 9.9% in 2006-2007 (807/ 8154). Conclusion: Of patients treated for secondary prevention of cardiovascular events with low-dose ASA, more than 50% received this as monotherapy. However, prescribing patterns for the prevention of these events have changed considerably during the past decade, with the proportion of patients receiving low-dose ASA in combination with clopidogrel and/or a PPI increasing over time. W1005 Natural History of Erosive Esophagitis, the Kalixanda Study, a Population- Based Study Jukka Ronkainen, Nick Talley, Tom Storskrubb, Michael Vieth, Tore Lind, Lars Agreus, Pertti Aro Background and aims: The natural history of gastroesophageal reflux disease (GERD) is under debate. The aim of this study was to determine the natural history of erosive esophagitis (EE) in a representative random sample of an adult Swedish population. Methods: A random sample (n= 1.000, mean age 53.5 years, 49% men) of the adult population (n= 21.610) in two Swedish municipalities was surveyed using a validated questionnaire assessing troublesome gastroesophageal symptoms and by esophagogastroduodenoscopy (EGD), response rate 73%. Those with EE according to LA-classification, endoscopically and/or histologically suspected Barrett's esophagus (BE), specialized intestinal metaplasia (SIM), intestinal metaplasia and every second of those with non-erosive reflux disease (NERD) were invited to a follow-up endoscopy five years after the primary investigation. During these years the subjects could consult and be treated as ordinary patients. Gastroesophageal reflux symptoms were defined as the presence of troublesome heartburn and/or acid regurgitation during the last three months in accordance with the Montreal classification. Confirmed BE was diagnosed when suspected columnar lined esophagus (sCLE) was detected with histologically diagnosed SIM. NERD was defined as the presence of troublesome reflux symptoms without EE or confirmed BE. Those with no reflux symptoms, EE and confirmed BE were defined as No-GERD. Results: Follow-up EGD with biopsy was performed in 289 subjects, response rate 71.0%. Symptom evaluation was available in 286 subjects, mean age 59.9 years, 54.6% men. Troublesome reflux symptoms were reported by 128 subjects (44.7%) while reflux esophagitis was found in 47 subjects (16.4%) and BE in 24 subjects (8.4%). The results, at the follow- up, of the 90 individuals who had had EE without BE at the primary investigation are shown in the table. Conclusion: EE may regress in 60%, a third remains constant and progression to BE can occur in about 9% within 5 years in a GERD-cohort of the general population. EE seems not to be a categorical disease. EE at five year follow-up in the general population S-630 AGA Abstracts * = some time during the follow-up, treated/untreated vs. controls ^n= 10 grade A, 10 grade B and 7 grade C “n=6 male **= positive/negative vs. controls W1006 Conjoint Analysis of Patient Preferences for 5ASA Maintenance Therapy in Ulcerative Colitis Joanna Dionne, Deborah A. Marshall, Lehana Thabane, John K. Marshall BACKGROUND:5-aminosalicylic acid (5ASA) is a first-line therapy for mild to moderately active ulcerative colitis (UC). Because non-compliance has a negative impact on long-term health outcomes, a conjoint analysis (CA) of patient preferences for 5ASA maintenance therapy was performed. CA is useful when deriving preference, as the principle behind this methodology is any product can be described based on features or attributes. It is these features that drive what product an individual will choose. METHODS: Structured focus groups and a systematic literature review were conducted to identify key attributes of 5ASA maintenance therapy important to patients: (1) effectiveness; (2) route of administration; (3) side effects; and (4) dose frequency. A discrete choice questionnaire with a full factorial design was constructed with 2 levels per attribute and 16 scenarios per respondent. Each choice task offered two alternatives (A or B) with the attributes shown in full profile to the respondent. The questionnaire was pilot tested with community members and UC patients to ensure readability and if attributes and levels elicited trading. Three differently-ordered versions of the questionnaire were administered to consecutive patients with UC attending outpatient gastroenterology clinics at the McMaster University Medical Centre. Disease sever- ity was assessed for each respondent with a partial Mayo Score. Responses were analyzed for main effects using Sawthooth Software SSI Web version 6.6 (Sequim, WA) using multino- mial logit estimation. RESULTS: Sixty UC patients completed the questionnaire (mean age 38.1 years, SD 14.6, range 17-78; 57% female). Mean age at diagnosis was 27.7 years (SD 13.9). The mean partial Mayo Score among 55 participants was 2.5 (SD 2.8) indicating mild disease. All participants used 5ASA therapy. All four attributes were statistically significant determinants of preference: Effectiveness was most important (β coefficient=13.95; p<0.05), followed by route of administration (11.61; p<0.05), side effects (6.23; p<0.05) and dose frequency (3.47; p<0.05) all at one degree of freedom. CONCLUSION: Patients place a high value on the effectiveness of therapy. Understanding which attributes of UC maintenance therapy are important to patients can help to design treatment strategies that improve both adherence and health outcomes. W1007 Aspirin and Nonsteroidal Anti-Inflammatory Drugs and the Risk of Diverticular Complications Lisa L. Strate, Yan Liu, Edward S. Huang, Edward Giovannucci, Andrew T. Chan BACKGROUND and AIMS: Non-steroidal anti-inflammatory drugs (NSAIDS) and aspirin may increase the risk of diverticular complications. We examined the associations between aspirin and NSAID use and the risk of diverticulitis and diverticular bleeding in a large prospective cohort. METHODS: We studied prospectively 47,210 US males in the Health Professionals Follow-up Study cohort who were aged 40-75 years and free of diverticular disease, inflammatory bowel disease, and cancer at baseline in 1986. Men with diverticulitis or diverticular bleeding were identified based on report of incident diverticulitis/ diverticulosis on biennial questionnaires, responses to a supplemental diverticular questionnaire, and beginning in 2006, report of bleeding from the colon requiring hospitalization or blood transfusion. Aspirin and NSAID use and other risk factors were assessed and updated biennially. We used Cox proportional hazards regression to calculate multivariate relative risks (RR) and 95% confidence intervals (CI) adjusting for age, study year, body mass index, dietary fat, fiber, red meat, nut, corn and total caloric intake, physical activity and concurrent aspirin or NSAID use. RESULTS: We identified 935 cases of diverticulitis and 256 cases of diverticular bleeding during 22 years of follow-up. Current regular aspirin users (≥2 times per week) had a multivariate RR of 1.13 (95% CI, 0.98-1.29) for diverticulitis and 1.45 (95% CI, 1.11-1.89) for bleeding when compared to non-regular users. The RRs of diverticulitis and bleeding increased with longer duration of use (P for trend 0.034 and 0.002, respectively). Compared with men who used no aspirin, the multivariate RRs for bleeding were 1.16 (95% CI, 0.71-1.89) for men who took 0.5-1.5 standard (325-mg) tablets per week, 2.07 (95% CI, 1.50-2.86) for 2-5 tablets per week, and 1.22 (95% CI, 0.80-1.87) for 6 or more tablets per week (P for trend 0.25). The RR for diverticulitis did not appear associated with increasing dose (P for trend 0.68). Current regular NSAID users (≥2 times per week) were also at increased risk for diverticular complications when compared to non-regular users: the multivariate RR was 1.45 (95% CI, 1.26-1.68) for diverticulitis and 1.36 (95% CI, 1.04- 1.79) for bleeding. Increasing duration of NSAID use was independently associated with the risk of diverticulitis (P for trend 0.001) but not diverticular bleeding (P for trend 0.07). CONCLUSIONS: Data from this large prospective cohort indicate that regular aspirin and NSAID use increase the risk of diverticular bleeding. Moreover, NSAID use is also associated with the risk of diverticulitis.